Inorganic Chemistry, Vol. 17, No. 3, I978 779
Correspondence it is not an excited-state association reaction as originally suggested.’ All attempts to develop a photochromic system which could be repeatedly cycled have failed. Experimental Section
Acknowledgmente We thank the National Science Foundation for support of this research.
Methyl isocyanide,” tert-butyl i ~ o c y a n i d e , ’and ~ [IrCI( 1,5~yclooctadiene)]~‘~ were prepared by published procedures, and solvents were purified by standard methods. All experiments and manipulations of compounds were conducted under vacuum or under a purified N2 atmosphere. Preparation of [Ir(CNCH3)4]Cl and [I~(CN-~-BU)~]CI. Excess C H 3 N C was distilled into a Schlenk tube containing [IrCI(I,Scycl~octadiene)]~ (0.30 g; 0.4 mmol) in 20 mL of CH2CI2,and the mixture was stirred for 1 h. The solvent, excess CNCH,, and 1,5-~yclooctadienewere removed under vacuum. The crude product was dissolved in methanol and filtered. Evaporation of solvent gave a blue-black solid. Other salts were prepared by metathesis in methanol using NaBPh4 and NaBF4. [ I ~ ( C N - ~ - B U ) ~was ] C Iobtained as an orange solid by an exactly analogous procedure using excess tert-butyl isocyanide. General Irradiation Procedures. Irradiations were conducted using a 450-W Hanovia medium-pressure Hg lamp in a quartz well equipped with one of the following Corning glass filters: 5-74 (A 436 nmj; 7-83 (A 366 nm); 3-70 (A 2525 nm). The complex to be studied was placed in an degassable quartz UV cell or a Schlenk tube, and after degassing on a vacuum line, the appropriate solvent was distilled onto the sample. Solutions for infrared studies were transferred in an inert-atmosphere glovebox to 0.5-mm NaC1-solution infrared cells. Lamp intensities were measured using ferrioxalate actinometryi6 and were of the order of 4.0 X IO-’ cinstein/min (436 nm).
References and Notes
Registry No. [Ir(CNCH3)4]C1, 40226-52-6; [Ir(GN-t-Bu),]CI, 34389-90-7; [IrC1(1,5-~yclooctadienej]~, 12112-67-3. W. M. Bedford and G. Rouschias, J . Chem. Soc., Chem. Commun., 1224 (1972). Similar notions have been suggested previ~usly.~*~ K. Kawakami, M.-A. Haga, and T.Tanaka, J . Orgunomet. Chem., BO, 363 (1973). J. W. Dart, M. K. Lloyd, R. Mason, and J. A. McCleverty, J . Chem. SOC.,Dalton Trans., 2039 (1973). Precedent for association of d8 complexes in solution H. Isci and W. R. Mason, Inorg. Chem., 13, 1175 (1974). K. R. Mann, J. G. Gordon, and H. B. Gray, J. Am. Chem. Soc., 97, 1551 (1975). \ - -
- I
G.-L. Geoffroy, M. S. Wrighton, G. S. Hammond, and H. B. Gray, J . Am. Chem. SOC.,96, 3105 (1974). R. Brady, B. R. Flynn, G. L. Geoffroy, H. B. Gray, J. Peone, Jr., and L. Vaska, Inorg. Chem., 15, 1485 (1976). G. L. Geoffroy, H. Isci, J. Litrenti, and W. R. Mason, Inorg. Chem., 16. 1950 . ., . ... (1977) A. H Reisi Jr.,V. S. Hagley, and S. W. Peterson, J. Am. Chem. Sac., 99 - , 4184 - - (1977)
A. P. Giniberg, Abstracts, New York Academy of Sciences Conference on Synthesis and Properties of low-Dimensional Materials, New York, N.Y., June 13-16, 1977, No. 21. R. E. Schuster, .I. E. Scott, and J. Casanova, Jr., “Organic Syntheses”, Collect. Vol. V, Wiley, New York, N.Y., 1973, p 772. G. W. Gokel, R. P. Widera, and W. P. Weber, Org. Synth., 55,96 (1976). J. L. Herde, J. C. Lambert, and C. V. Senoff, Inorg. Synth., 15, 18 (1974). C. G. Hatchard and C. A. Parker, Proc. R . SOC.London, Ser. A, 235, 518 (1956).
1
Correspondence Reinvestigation of the Coordination Geometry of Eight-Coordinate Metal Tetrakis(acety1acetonates)
Table 1. Shape Parameters (deg) for M(acac), Complexes‘
Sir:
D2d dodecahedron*
Although several sets of parameters have been devised to measure the shape of a discrete eight-coordinate complex with reference to idealized geometric forms,I4 some confusion still exists in the assignment of an idealized geometry to an observed eight-coordinate complex. A case in point is the series of acetylacetonate (acac) complexes with actinides, lanthanides, and group 4B elements. The individual eight-coordinate tetrakis(acety1acetonate) complexes exist in either of two crystalline modifications (the a and /3 forms); the coordination geometry of the 0 form has been described as primarily square antiprismatic,5-s while that of the a form has been described as predominantly dodecahedra19J0 and as predominantly antiprismatic.]’ We have reexamined the tetrakis(acety1acetonate)series by calculating the 4 and 6 shape parameter~~-~,’* for a-Ce(acac)4, a - T h ( a ~ a c ) ~p-Zr(acac),, , B-Ce(acac),, p-U(acac),, and pNp(acacI4. The results are given in Table I. The results of mean-planes calculations, assuming both dodecahedral and antiprismatic geometries for both series of complexes, are listed in Table 11. Tables 111and IV give the atomic coordinates and the definition of the planes used in the calculation of the shape parameters for the a and 0 forms, respectively. The idealized geometry most closely approximated by the coordination groups of a - C e ( a ~ a cand ) ~ a-Th(acac)4 is neither the D2d dodecahedron nor the DM square antiprism but rather the C,, bicapped trigonal prism, as shown by the shape parameters in Table I. The assignment of primarily dodecahedral geometry to the coordination group of the a complexes by Allard9 was based on a comparison of observed normalized polyhedral edge lengths with the corresponding edge lengths 0020-1669/78/ 1317-0779$01 .OO/O
Compd
(6,
0.0 C,, bicapped trigonalprism* 14.1 D, square antiprismb 24.5 a-Ce(acac) 12.8 orTh(acac), 15.6 p-Zr(acac), 22.8 pCe(acac), 19.7 @-U(acac), 22.1 22.3 BNp(acac),
(62
6,
6,
6 3
64
0.0 29.5 29.5 29.5 29.5 14.1 0.0 21.8 48.2 48.2 24.5 0.0 0.0 52.4 52.4 15.3 9.7 21.1 44.0 42.9 17.3 5.4 19.8 44.5 44.9 22.4 3.2 3.2 49.6 49.6 19.3 5.3 5.3 46.7 46.7 17.8 6.8 6.8 49.2 49.2 20.6 4.5 4.5 50.5 50.5
a The planes defiiing the rp and 6 angles are given for the a complexes in Table I11 and for the p complexes in Table IV. Values of the shape parameters for the idealized geometries were taken from ref 4.
for the “most favorable polyhedra”.’ However, it is very difficult on this basis to assign an idealized geometry to the a complexes since their edge lengths are approximately midway between those of the dodecahedron and the square antiprism. If anything, there appears to be a slight tendency toward antiprismatic coordination. Our calculation of the 4 and 6 shape parameters assigns the geometry of the a complexes as quite close to bicapped trigonal prismatic. Moreover, since the bicapped trigonal prism is along the reaction pathway between the dodecahedron and square a n t i p r i ~ m ,our ~ ? ~calculation is in quite good agreement with the polyhedral edge lengths given by Allard. The mean-planes calculations for a-Ce(acac)4 and a-Th(acac), (see Table 11) show a large degree of nonplanarity for the BAAB trapezoids, again indicating that the a complexes do not closely approximate dodecahedral geometry. One of the square faces for the antiprismatic model is also relatively nonplanar. Note, however, that the deviations of the atoms from the plane of the other square face, the face corresponding 0 1978 American Chemical Society
780 Inorganic Chemistry, Vol. 17, No. 3, 1978
Correspondence
Table 11. Mean Planes for M(acac), Comolexes
BAAB trapezoids
Square facesa
Compd
Plane 1
Plane 2
Plane 1
Plane 2
a-Ce(acac),
0.141b -0.224 0.224 -0.141 0.167 -0.270 0.287 -0.1 84 -0,236 0.376 -0.376 0.236 -0.229 0.345 -0.345 0.229 -0.259 0.398 -0.398 0.259 -0.245 0.396 -0.396 0.245
- 0.17 2b 0.269 -0.266 0.169 -0.202 0.312 -0.304 0.194 0.234 -0.367 0.367 -0.234 -0.222 0.344 -0.344 0.222 -0.193 0.310 --0.3 10 0.193 .- 0.23 2 0.358 -0.358 0.232
0.086b -0.084 0.083 -0.085 0.047 -0.048 0.049 -0.047 0.027 -0.027 -0.027 0.027 0.046 -0.046 - 0.046 0.046 0.057 -0.057 -0.060 0.060 0.038 -. 0.037 -0.039 0.038
0.186b -0.189 0.193 -0.190 0.180 -0.183 0.182 -0.179
a-Th(acac),
p-Zr(acac),
0-Ce(acac),
0-U(acac),
P-NP (acac) ,
Dihedral angle between trapezoids, de L
Dihedral angle between square faces, deg
88.6
1.6
90.0
1.1
84.1
0.0
85.5
0.6
85.1
3.5
83.4
0.0
a In the p crystalline form the two square faces are rendered equivalent by a crystallographic twofold axis, displacements (A) of the individual atoms from the planes defined in Tables 111 and KV.
The numbers given are the
Table III. Atomic Coordinates" and Shape Parameter Definitions for a-M(acac), Complexes Atom
Shape parameterb
X
Y
2
0.1903 0.0618 0.0398 0.2108 0.0098 0.1888 0.3885 0.3010 0.3297
0.1464 0.1791 0.0150 0.3323 0.22 19 0.1200 0.19 14 -0.0141 0.1361
0.2003 0.0562 0.1535 0.2005 0.2061 0.3379 0.3011 0.2275 0.1256
Plane 1
"3 Figure 1. A view (ORTEP)of the coordination group of ac-Th(acac)4 showing its bicapped trigonal prismatic shape.
Plane 2
@l
@a 61
Sa 83 84
BAAB trapezoidsd Square faces a The fractional atomic coordinates used in the calculations were taken from ref 10 and 11. The coordinates shown are those for a-Th(acac),. The @ and 6 parameters are dihedral angles between plane 1 and plane 2. refers to the midpoint between oxygen atoms x and y . '??le trapezoids given are those for the gggg stereoisomer.
to the noncapped square face of the bicapped trigonal prism, are rather small. A view of the coordination group in a-Th(acac)4 emphasizing its bicapped-trigonal-prismatic nature is shown in Figure I , and a view down the pseudo-threefold axis of the trigonal prism is presented in Figure 2. Figure 3 shows the ligand wrapping pattern for a-Th(acac)4. Oxygen atoms 1, 3. and 8 and 2, 5 , and 7 form the triangular faces of the trigonal prism: oxygen atoms 4 and 6 cap two of the three quadrilateral
Figure 2. A view (ORTEP)down the pseudo-threefold axis of the coordination group of a-Th(acac)4.
K-1 Figure 3. A view (ORTEP)down the pseudo-8 axis of a-Th(acac)4.
Inorganic Chemistry, Vol. 17, No. 3, I978 781
Correspondence Table IV. Atomic Coordinatesu and Shape Parameter Definitions for p-M(acac), Complexes Atom
X
Y
NP
0.0000 0.0623 0.1121 0.0507 0.0089 -0.0623 -0.1121 -0.0507 -0.0089
-0.0676 -0.2821 0.0186 -0.1548 0.1426 -0.2821 0.0186 0.1 548 0.1426
0,
0, 0 3 0 4
0 1'
0,' 0,: 0,
Shaoe oarameterb
6, 6,
2
-
Plane 2
Plane 1
O,O,'O,'
030402'
O,'O,'O,'
O,O,'O,' 0,'O3'O,'
o,o,o,'
63
6,
BAAB trapezoidsd Square faces
0.2500 0.2410 0.3025 0.4175 0.355 1 0.2590 0.1975 0.0825 0.1449
010~03 O,'O,'O 0
o,o,o,'~,?
07.0304
0,'0,'0,0,
e
The fractional atomic coordinates used in the calculations were taken from ref 5-8. The coordinates shown are those The @ and 6 parameters are dihedral angles for O-Np(acac),. Mxyrefers to the midpoint bek t w e e n plane 1 and plane 2. The trapezoids given are those tween oxygen atoms x and y . for the m m m m stereoisomer. e The two square faces are rendered equivalent by a crystallographic twofold axis.
Table V. Polyhedral Edge Lengths for a-M(acac), Complexes Length, A Edgeu
t, t, p1
p, h,
h, a
Atoms 0,-0, 0,-0, 0,-0, 0,-0, 0,-0, 0,-0,
O,-O, 03-0,
O,-O,
0,-0, 0,-0,
0,-0, 03-05
0,-0, 06-0, 0,-0,
O,-O,
orCe(acac), 2.803 3.239 3.009 2.747 2.817 2.743 2.790 2.713 3.656
2.736 3.251 2.991 2.883 2.847 2.775 2.731 2.731
a-Th(acac), 2.871 3.253 3.179 2.859 2.895 2.751 2.787 2.747 3.658
2.970 3.352 3.114 2.891 2.927 2.801 2.805 2.695
See Figure 1and ref 3 for a defmition of edge nomenclature.
faces. The observed ligand wrapping pattern for the a-M(acac), complexes gives the h l h l p g 2stereoisomer of the bicapped trigonal prism according to the notation of PoraiKoshits and as la no^.^^'^ The h l h l p g 2stereoisomer lies along the deformation pathway between the ssss square antiprismatic stereoisomer and the gggg dodecahedral stereoisomer. If the ssss stereoisomer (see Figure 3) is taken as the starting point, then a folding of one square face about the 030,diagonal and a slight shortening of the 0305 diagonal produce the observed h l h l p g 2bicapped trigonal prismatic stereoisomer. A folding about and slight shortening of the 0208 diagonal of the remaining square face would then lead to the gggg dodecahedral stereoisomer. It is therefore easy to understand how the coordination geometry of the a-M(acac), complexes could have been interpreted as ssss antiprismatic or gggg dodecahedral. The polyhedral edge lengths for a-Ce(acac), and a-Th(acac), are listed in Table V. There are some significant deviations from the edge lengths of the idealized bicapped trigonal prism based on the hard-sphere model. For example, the two types of edges ( t l and t 2 )which define the triangular faces of the trigonal prism differ in length, and the longer t2 edges are also longer than the edges @, and p 2 ) which terminate at the capping atoms; all four edges (t,, t2,pl, and p 2 ) are required to be identical in the hard-sphere model. The p1 and p 2 edges are, however, remarkably uniform with an average length of 2.792 8, and a range of 2.731-2.883 A for the Ce complex and an average length of 2.840 8, and a range
Figure 4. A view (ORTEP) down the pseudo-8 axis of P-Np(acac),. The primed and unprimed oxygen atoms are related by a crystallographic twofold axis.
of 2.757-2.927 8, for the T h complex. Thus, the capping atoms are very nearly centered on the quadrilateral faces. The 0102 and 0708 edges ( h , ) are spanned by the bidentate acetylacetonate ligands and are therefore much shorter than edge (h2). The two triangular faces of the trigonal the 0305 prism are thus nonparallel (dihedral angle, 19.1' for a-Ce(acac), and 18.7' for a-Th(acac)d, and the capped faces are trapezoidal rather than rectangular. A comparison of Figures 1-3 and the view of an a-M(acac), complex given by Allard9 shows that computer-drawn molecular diagrams can sometimes be misleading. In the view given in Allard's paper the molecule of a-M(acac), appears to approximate dodecahedral geometry while in the view shown in Figure 3 it appears to closely approximate square-antiprismatic geometry. However, our calculations, as discussed above, indicate that the a-M(acac), complexes most closely approximate bicapped-trigonal-prismatic geometry, as shown in Figures 1 and 2. Therefore, when a series of possible geometries are closely related and lie along a reaction pathway, such as in the case of eight-coordination, tabulations of 6 and I$ shape parameters, mean planes, polyhedral edge lengths, and interbond angles should be given in conjunction with the chosen view to accurately depict the geometry of the complex. The assignment of square-antiprismatic geometry to the P-M(acac)4 complexes is unambiguous. Our calculation of the 6 and I$ shape parameters (see Table I), as well as Allard's listing of edge lengths, indicates that the square antiprism is the idealized geometry closest to that of the P-M(acac), complexes. The slight distortions from idealized squareantiprismatic geometry involve folding (see Figure 4) about the 0302/and 0203/diagonals, distortions in the direction of the mmmm dodecahedral ~ t e r e o i s o m e r . ~ * ~ A further example of ambiguity in the assignment of coordination geometry involves tetrakis(N,N-diethyldithiocarbamato)thorium(IV), Th(Et,dtc),, an eight-coordinate complex whose geometry has been described as intermediate between dodecahedral and square antipri~matic.'~The assignment of intermediate geometry was based on a calculation of OA and OB, where 0 A and OB are the angles which the M-A and M-B bonds make with the 4 axis of the D2ddodecahedron or with the axis of the D4dsquare antiprism.' The values for the "most favorable polyhedra" are OA = 35.2' and OB = 73.5' for the dodecahedron and 6~ = 6, = 57.3' for the square antiprism.l The reported values of BA = 44' and OB = 66' for the thorium complex14 are, however, the angles which the M-A and M-B bonds make with a crystallographic twofold axis which is not coincident with the 3 axis of the dodecahedron but rather perpendicular to it. A calculation of the angles which the M-A and M-B bonds make with the 3 axis of the dodecahedron (taken as the line passing through the midpoints of the two a edges and through the thorium atom) yields values of 34.8' for OA and 83.0' for OB. Thus, Th(Et2dtc), is apparently much closer to dodecahedral (mmmm stereoisomer) than to square-antiprismatic geometry. A determination of
782 Inorganic Chemistry, Vol. 17, No. 3, 1978
Correspondence
the 6 and r$ shape parameters for Th(Et2dtc), confirms its approximation to dodecahedral geometry; the 6 parameters are 31.4, 33.5, 41.9, and 41.9’ and the one unique 4 value is 5.1’. (See Table I for a listing of the 6 and 4 parameters for the idealized dodecahedron and square antiprism.) The dihedral angle between the interpenetrating BAAB trapezoids of the dodecahedron is 89.6’, with the four sulfur atoms which define an individual trapezoid being planar to within 0.1 1 A. The cases of misdescription of coordination geometry discussed above emphasize that the idealized eight-coordinate geometries are closely related by fairly small deformations along a reaction pathway and that assignment of one of the idealized geometries to an observed complex should be based on an analysis of more than one set of shape parameters. Although calculations of the eA and OB angles and of normalized edge lengths are often very useful in describing the geometry of a complex, a tabulation of 6 and r$ shape parameters, as well as results of appropriate mean-planes calculations, should be provided in defense of an assigned geometry. In addition, in cases of tetrakis complexes with bidentate chelating ligands, the ligand wrapping pattern should be identified. A clear and convincing description of the geometry of an observed complex can thus be given. Acknowledgment. The support of this research by National Science Foundation Grant CHE-7620300 is gratefully acknowledged.
(2) S. J. Lippard and N. J. Russ, lnorg. Chem., 7. 1686 (1968). (3) M. A. Porai-Koshits and L. A. Aslanov, J . Struct. Chem. (Engl. Transl.), 13, 244 (1972). (4) E. L. Muetterties and L. J. Guggenberger, J . A m . Chem. Soc., 96, 1748 (1974). ( 5 ) J. V. Silverton and J. L. Hoard, Inorg. Chem., 2, 243 (1963). (6) H. Titze, Acta Chem. Scand., 23, 399 (1969). (7) H. Titze, Acta Chern. Scand., 24, 405 (1970). (8) B. Allard, Acta Chem. Scand., 26, 3492 (1972). (9) B. Allard, J. Inorg. Nucl. Chem., 38, 2109 (1976). (10) B. Allard, Acta Chem. Scand., Ser. A , 35, 461 (1976). (11) H. Titze, Acta Chern. Scand., Ser. A , 28, 1079 (1974). (12) The four 6 shape parameters are the dihedral angles between the pairs of triangular faces which join along the four 6 edges of the reference D2d dodecahedron. These four dihedral angles are chosen as a criterion of shape because they are more sensitive to polyhedral shape than the dihedral angles between triangular faces which join along the (I, m , and g edges of the reference dodecahedron. Upon conversion to the D, square antiprism, a pair of opposite dodecahedral 6 edges become diagonals of the square faces of the antiprism, and the corresponding 6 values decrease from 29.5’ in the idealized dodecahedron to 0.0’ in the idealized square antiprism; the other two dodecahedral 6 edges become 1 edges of the antiprism and the corresponding 6 values increase to 52.4’ (see Table I). The C2”bicapped trigonal prism has one square face and one 6 value equal to 0.0’. Identification of the shape-determining edges (6 edges of the reference dodecahedron) is discussed in ref 4. (1 3) It is interesting to note that Porai-Koshits and Aslanov excluded from , trigonal prism consideration as a “main isomer” the h l h l p g pbicapped for complexes with four identical bidentate ligands, on the basis of the difficulty of identical ligands spanning edges which have different lengths in the idealired hard-sphere polyhedron. (14) D. Brown, D. G . Holah, and C E. F. Rickard, J . Chern SOC.A, 423 (1970)
Registry No. a-Ce(acac),, 6 5 137-07-7; a-Th(acac)4, 65 137-06-6; p-Zr(acac),, 65137-05-5; p-Ce(acac),, 65137-04-4; p-U(acac),, 65137-03-3; P-Np(acac),, 65137-02-2.
Department of Chemistry Cornell University Ithaca. New York 14853
References and Notes (1) J. L. Hoard and J. V. Silverton, Inorg Chem., 2, 235 (1963).
Receiced June 28, 1977
William L. Steffen Robert C . Fay*
